Magnet resonance imaging has shown a remarkable sensitivity to blood flow. We propose to use this totally noninvasive modality, in a unique way, to provide angiographic images of the venous system. In general, magnetic resonance provides cross-sectional images of the anatomy which have little value as an angiographic tool for finding thrombi in veins. These cross-sectional images usually image flowing material as a void, or lack of signal. Under some circumstances the blood become enhanced when it represents freshly excited spins compared to the partially saturated spins of the surrounding static tissue. The Stanford medical imaging group first introduced the concept of projection vessel imaging using magnetic resonance to obtain angiographic presentation. In this context static material is cancelled, and blood is imaged as an active source. Three basic mechanisms can be used: time of flight, where excitation and reception take place at different positions; substraction, where images representing different velocities or velocity sensitivities are subtracted; and phase shift, where blood experiences a phase shift different from that of static tissue. Many approaches represent a variety of these general themes. We have recently implemented systems which, following verification in phantom and human studies, appear to completely meet the stated requirements of venous imaging. These systems can produce high SNR, high resolution images with inherent quantitative velocity information. Arbitrarily large field-of-view images of the venous system are also possible. The methods offer sufficient flexibility to provide all of the required tradeoffs in performance. We plan to fully study its application to veins in both normal and diseased humans. We will also study the application of our recently developed spectroscopic imaging methods to help quantify the state of the diseased vessel.